Consistency of inconsistency in long-COVID-19 pain symptoms persistency: A systematic review and meta-analysis

Oleg Kerzhner MD, Einat Berla MD, Meirav Har-Even MSc RD, Motti Ratmansky MD, Itay Goor-Aryeh MD, July 2023, https://doi.org/10.1111/papr.13277

Introduction

Individuals recovering from acute COVID-19 episodes may continue to suffer from various ongoing symptoms, collectively referred to as Long-COVID. Long-term pain symptoms are amongst the most common and clinically significant symptoms to be reported for this post-COVID-19 syndrome.

Objectives

This systematic review and meta-analysis aimed to evaluate the proportions of persisting pain symptoms experienced by individuals past the acute phase of COVID-19 and to identify their associated functional consequences and inflammatory correlates.

Methods

Two online databases were systematically searched from their inception until 31 March 2022. We searched primary research articles in English, which evaluated individuals after laboratory-confirmed COVID-19 acute phase resolution and specifically reported on pain symptoms and their inflammatory and/or functional outcomes.

Results

Of the 611 identified articles, 26 were included, used for data extraction, and assessed for their methodological quality and risk of bias by two independent reviewers. Pain symptoms were grouped under one of six major pain domains, serving as our primary co-outcomes. Proportional meta-analyses of pooled logit-transformed values of single proportions were performed using the random-effects-restricted maximum-likelihood model. An estimated 8%, 6%, 18%, 18%, 17%, and 12% of individuals continued to report the persistence of chest, gastrointestinal, musculoskeletal joint, musculoskeletal muscle, general body, and nervous system-related pain symptoms, respectively, for up to one year after acute phase resolution of COVID-19. Considerable levels of heterogeneity were demonstrated across all results. Functional and quality-of-life impairments and some inflammatory biomarker elevations were associated with the persistence of long-COVID pain symptoms.

Conclusion

This study’s findings suggest that although not well characterized, long-COVID pain symptoms are being experienced by non-negligible proportions of those recovering from acute COVID-19 episodes, thus highlighting the importance of future research efforts to focus on this aspect.

INTRODUCTION

While the first cases of coronavirus disease 2019 (COVID-19) began to rise above the surface towards the end of 2019,¹ their exponential surge by early 2020 left no room to question the existence of an unfolding global pandemic, which was led by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) agent.² According to the “World Health Statistics 2022” report, issued by the world health organization (WHO), 504.4 million confirmed cases of COVID-19, with 6.2 million deaths directly attributed to it, were reported until the 20th of April 2022. While these figures probably underestimate the true case-positive rate, as roughly 75% of reports arrived from developed WHO regions,³ they are nonetheless well indicative of the overall associated health and socioeconomic burden.^4–7

Acute COVID-19 cases may be subclinical, or present with various symptoms, ranging from mild fatigue, cough, fever, or anosmia to most intense and fatal cases such as overt respiratory failure and multiorgan collapse.^8–10 This symptomatic diversity soundly reflects the multisystemic nature of COVID-19, which may involve almost any organ or system of the human body.^10, 11 For the most part, acute COVID-19 episodes are self-limited, and most individuals achieve full recovery.¹² However, as high as 80% of patients may continue to experience ongoing symptoms for weeks to indefinitely past the acute phase resolution.^13, 14 These cases constitute the postacute sequelae of SARS-CoV-2 infection (PASC) and seem to be independent of acute disease severity and temporal continuity, thus posing a challenge to determine whether an individual’s set of ongoing symptoms represents normally prolonged recuperation or are part of long-COVID syndrome (LCS).^13, 15, 16 Moreover, a substantial level of uncertainty concerning various clinical aspects of these long-term disease effects still exists.⁸ PASC, long-COVID, or post-COVID-19 syndrome (PCS) are just a few examples of the interchangeable terminology used to describe this condition, as consensus definitions are yet to be universally adopted.^17–20 Similarly, considerable variability exists in the minimal time periods used for determining long-COVID cases.^{17, 18, 20–22} Cutoff periods between two weeks and up to 24 weeks postacute phase disease laboratory confirmation or hospital discharge, were reported to be used for that purpose.^13, 18, 20

Recognizing the barriers to better epidemiological characterization and treatment development posed by these issues has led to some attempts to establish common grounds for a globally accepted framework.³ For example, several European medical societies formulated LCS guidelines, aimed at providing more efficient identification, evaluation, and mitigation of COVID-19’s persistent effects.^21, 23 The National Institutes of Health (NIH) proposed the term PASC to be used when symptoms persist for at least 28 days after acute infection.²⁴ A recent WHO-led international Delphi process determined that PCS designation should follow at least 2 months of symptoms and cannot be explained by alternative diagnosis in individuals with probable or confirmed SARS-CoV-2 infection.²² However, with constantly emerging new evidence and increasing understanding concerning the long-term effects of COVID-19, it is acknowledged that these definitions are still evolving and thus would have to be refined and updated in the future.^22, 24

These taxonomical issues could also be attributed to the highly diverse clinical spectrum characterizing the globally reported cases of long-COVID.^13, 22 Among the most frequently reported symptoms are chest pain, fatigue, dyspnea, and cough, with prevalence rates of up to 89%, 65%, 61%, and 59%, respectively.¹³ These, together with other not uncommonly reported symptoms such as joint pain, muscle aches, headache, forgetfulness, asthenia, anxiety, depression, dysphagia, anosmia, and ageusia, constitute merely a fraction of the entire repertoire of long-COVID-19 putative clinical symptomatology.^{13, 16, 24, 25} This clinical heterogeneity presents another barrier in the global endeavor to facilitate appropriate levels of accurate diagnosis and overall management of this condition.^14, 23

Pain symptoms persistence past the acute phase resolution of COVID-19 is repeatedly cited as one of the most common long-COVID-19 symptoms to be reported,^11, 26, 27 with some reports holding them as the most common.²⁸ Nonetheless, the vacillating nature of long COVID-19 demonstrates high pervasiveness even when attempting to assess more granular levels of clinical representation. As such, pain symptoms in long-COVID are no exception to the rule, as wide ranges of prevalence were reported using mostly nonstandardized assessment instruments together with inconsistent definitions concerning various clinical aspects of involved body systems, organs, and anatomical locations to suffer from pain.^{13, 16, 25, 29}

Although the reported data regarding various pain symptoms in long COVID-19 is constantly being updated, it still suffers greatly from methodological flaws due to the just mentioned issues as well as from relative research paucity and considerable knowledge gaps, even when compared to general long-COVID symptoms.^30–35 It should also be noted that these issues are of mutual concern for both pain and other nonlife-threatening long-COVID-19 symptoms.³⁴ These, in turn, undermine the ability to structure firm evidence-based knowledge to better guide the overall management of the various pain conditions.^13, 30, 31 Thus, despite being commonly reported, long-COVID-19 pain symptoms are not adequately described and require more extensive and in-depth definitions of their characterizing clinical features.^31, 36 A better understanding of the prevalence, diversity, and extent of these usually nonfatal but quality of life compromising long-COVID manifestations is, therefore, necessary for supporting the development of appropriate follow-up and intervention strategies for postacute phases of COVID-19.³⁴

Hence, in this study, we aimed to explore and gain new insights concerning the proportion of individuals suffering from persistent long-term pain symptoms in COVID-19 survivors. Functional outcomes, quality of life (QoL) measures, and inflammatory correlates of the various pain symptoms served as secondary outcomes.

METHODS

Literature search and reporting

A protocol was formulated in accordance with the statement of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.³⁷ The review protocol was not registered. Systematic search of English language literature was conducted on PubMed/MEDLINE and EMBASE, from database inception to the 31st of March 2022. Keywords used included “COVID-19” OR “coronavirus COVID-19” OR “COVID syndrome” OR “post-COVID syndrome” OR “long-COVID syndrome” OR “pain symptoms” OR “long-COVID syndrome symptoms” OR “post-COVID-19 sequelae” OR “postacute sequelae” OR “long-COVID syndrome pain” OR “postacute sequelae of SARS-CoV-2”. Two review authors (OK and EB) independently screened for titles and abstracts, retrieved relevantly identified articles for full-text assessment, and removed duplicates. Discrepancies were resolved through discussion.

Guidelines and selection criteria

We searched articles reporting on the incidence of any pain symptom (primary outcomes) and/or secondary outcomes related to them (eg, functional outcomes or inflammatory markers) in individuals who recovered from the acute phase of confirmed COVID-19. No restrictions were made to require a specific minimal time frame since diagnosis or recovery from the acute phase. Initially we aimed to investigate the secondary outcomes only for their correlation with long-COVID’s pain symptoms. However, due to the paucity of data concerning the latter we have decided to include secondary outcomes associated with long-COVID in general, and not pain specifically (Table 1). Any study that reported the persistence of pain symptoms past the acute phase of COVID-19 was considered as long as passing our inclusion and exclusion criteria.TABLE 1. Definition of study variables.

Domain and symptom definitions	Terms used for describing specific symptoms	Major domain definitions & considerations for inclusion or exclusion
Co-primary pain outcomes
Chest pain	Chest pain; Rib pain^b; Thorax-chest pain^a; Angina pectoris^a	Any pain symptoms reported to involve any part of the thoracic region^c^,^d^,^e^,^f
Gastrointestinal system-related pain	Dysphagia; Odynophagia; Abdominal pain; Sore throat; Sore throat or difficulty to swallow; Stomach pains^a	Any pain symptoms reported to involve any part of the GI tract and abdominal region^c^,^d^,^e^,^fSore throat was considered to represent the GI system unless specific information was provided to exclude oropharyngeal origin
Musculoskeletal system pain—Joints	Arthritis; Joint pain; Arthralgia	Any pain symptoms reported to involve or relate to joints^c^,^d^,^e^,^fIf general terms of joint pain (eg, Arthromyalgia) included also muscular components without further specifying which symptoms fall under this term, then it was excluded and considered musculoskeletal pain of muscle origin
Musculoskeletal system pain—Muscles	Arthromyalgia^a; Musculoskeletal pain; Myalgia; Spine pain^b; Muscle pain	Any pain symptoms reported to involve or relate to specific muscular structures^c^,^d^,^e^,^fIf general terms of muscle pain (eg, Muscle pain, Musculoskeletal pain, Myalgia) were reported without providing information about which exact anatomical regions or symptoms fall under this term and/or study did not provided separate (additional) category for symptoms eligible for the “General body pain” domain, then it was excluded and considered as “General body pain” symptom. Otherwise, it was included in this domain.
General body pain	Pain^a; Pain other than chest^a: Body pain; Body aches^a; Widespread pain^a	Any general term indicating wide and nonspecific distribution pattern of pain symptoms, without providing further specification about exact anatomical regions or symptoms and/or study providing separate category for symptoms eligible for the “Musculoskeletal system pain—muscles” domain than it was included in this domain^c^,^d^,^e^,^f
Nervous system-related pain	Headache; Tension type like headache; Migraine-like headache; Facial headache^a	Any pain symptoms reported to involve or relate to specific nervous system structures or involve any part of the head region apart from those regions that were prespecified to fall under any of the other domains (eg, oral region pain)^c^,^d^,^e^,^f
Not falling under any definition	Pain swelling or discharge of the eye^b

Mode of ascertainment	Objective	Subjective
Co-primary pain outcomes	Pain confirmed using anyvalidated tool (eg, the Self-reported symptoms questionnaire SRSQ)	Self-report or nonvalidated instrument for pain assessment (eg, general symptoms questioners used by the authors, self-reported symptoms by the patients during clinical follow-up).
Secondary outcomes
Inflammatory Parameters	Any measures providing levels for biomarkers ascertained by laboratory tests (for example including but not limited to such as CRP, IL-6, Neutrophil count, Lymphocyte count, D-dimer, and fibrinogen)	N/A
Functional outcomes/Quality of Life (QoL)	Functional impairment evaluated by using validated tools for assessment of QoL or functional outcomes (eg, EQ-5D-5L, 6MWT, SF-36, mMRC dyspnea scale).	Self-report or nonvalidated instruments for evaluation of functional impairment or general QoL status (eg, nonstandardized Likert scale questionnaires for assessing these domains, items in general questioners concerning issues of limitations in activity, work, social life, and general life satisfaction).

Abbreviations: 6MWT, 6-min walk test; CRP, C-reactive protein; EQ-5D-5L, EuroQol-5 Dimension-5 levels; IL, interleukin; mMRC, Modified Medical Research Council scale; N/A, not applicable; SF-36, 36-Item Short Form Survey; SRSQ, St. George’s respiratory questionnaire.
^a Indicates that a term was used only once.
^b Indicates that a term was used only once but was not considered for data synthesis after failing to qualify the definitions.
^c No system-based distinction was made for a given anatomic region, and if no specific information was provided to complement the anatomical description, then pain could thus possibly be referred from or related to the cardiac and/or pulmonary and/or musculoskeletal and/or nervous systems.
^d If incidence for two or more pain symptoms under this domain were reported, without providing information about the level of overlap between them (intrapatient symptom co-occurrence) then only the largest group was entered into synthesis while it was assumed that a maximal overlap exists between the various reported symptoms.
^e Pain symptoms that were reported to be present before COVID-19 acute phase without experiencing any worsening or exacerbation in their severity were not included.
^f If more than one follow-up period was provided for the same pain symptom, then the longest time period available was chosen to represent the reported incidence of pain symptoms.

Inclusion criteria:

Individual, of any age, sex, and ethnicity, with a confirmed diagnosis of COVID-19 (of any clinical severity) as confirmed through standardized assays (ie, laboratory testing of PCR, serological, or antibody biomarkers) and/or documented equivalent (eg, ICD-codes or electronic health records indicating confirmed COVID-19 cases).
Individuals reported by the authors as those with complete recovery from the acute phase COVID-19. Temporal restriction was not posed on reported follow-up duration since confirmation of acute phase COVID-19, and all studies reporting about recovering individuals past the acute phase of the disease, were included.
Studies reporting matched individuals who were either nonexposed to COVID-19 or reporting the status of individuals with COVID-19 prior to their acute infection. These comparators served our quality assessment of the component studies, as the MA was concerned with the prevalence among confirmed cases.
Full estimates of quantitative data, exact proportions (or appropriate quantitative equivalent of complete values), concerning primary outcomes, and/or quantitative/qualitative data concerning the secondary outcomes for acute phase, confirmed COVID-19 survivors of any demographic, as specified in Table 1.
Full-text published primary research articles.
Prospective or retrospective cohort, cross-sectional, and case-series studies. Case–control designs were included only if the long-COVID cohort was reported independently enough to serve as a separate cohort of general long-COVID with a clear incidence of our outcomes.

Exclusion criteria:

Selected cohort representing outcome population (eg, long-COVID patients suffering from odynophagia).
Incomplete or nonclear description of primary outcomes.
Outcomes reported before acute COVID-19 recovery.
Nonprimary research, abstracts, case reports, and studies with a cohort of < 20 individuals.
Provided information concerns only new symptoms arising following the resolution of acute phase COVID-19 while not addressing persistent symptoms since the diagnosis.

Data extraction and outcome measures

Data from studies that passed the exclusion and inclusion criteria were extracted independently by two reviewers (OK and EB) into a predefined extraction table, then corroborated, with discrepancies resolved through discussion. Extracted data included study design, country, sample source and characteristics, follow-up duration, ascertainment of COVID-19 and outcomes, outcomes (primary and secondary), associated factors with long-COVID, and terms used to describe COVID-19’s long-term effects (Table 2).TABLE 2. Characteristics and results of studies (n = 26) examining individuals with confirmed COVID-19 after recovery from the acute phase of the disease.

Study	Country	Study design	Sample source	Sample characteristics	Follow-up duration	Ascertainment of COVID-19	Ascertainment of outcomes	Outcomes N (%)	Factors associated with persistent symptoms	Terms used to refer to long-term effects
							● Primary	● Primary
							○ Secondary	○ Secondary
							◘ Other relevant	○ Secondary
Asadi-Pooya et al., 2021¹⁸	Iran	Retrospective observational study	55 centers throughout Fars province	N_total = 4681 (Total cohort; 2 groups separated by follow-up duration since hospital discharge, 3–6 and 6–12 months)▫ Age Range: ≥ 18▫ Mean Age: 52 ± 15▫ Males: 2478 (52.9%)▫ Females: 2203 (47.1%)N_{6–12 month} = 1996 (6–12-month follow-up cohort)	3–6 and 6–12 months after hospital discharge.	RT-PCR	● Subjective self-report via telephone conversation.○ Nonstandardized 5-item questionnaire based on a Likert scale^a	(For N_{6–12 month})● Muscle pain 291 (15%)● Joint pain 296 (15%)● Chest pain 175 (9%)● Sore throat 74 (4%)● Headache 207 (10%)● Abdominal pain 56 (3%)○ LCS is significantly associated with impaired ADLs and QoL of the patients	Female sex, severity of initial respiratory symptoms, and prolonged hospitalization were significantly associated with experiencing LCS	Long-COVID syndrome (LCS), Long-term symptoms associated with COVID-19
Aydin et al., 2021³⁸	Turkey	Cross-sectional study	Ataturk University, Education and Research Hospital	N = 116▫ Age Range: ≥ 18▫ Mean Age: 48.9 ± 17.74▫ Males: 56 (48.3%)▫ Females: 60 (51.7%)	At least 4 weeks from the initial diagnosis.	RT-PCR	● Subjective self-report (during clinical follow-up)○ Objective assessment via laboratory tests◘ CT severity score^b	● Chest pain 7 (6%)● Headache 11 (9.5%)	Persistence of chest pain was correlated with higher CT severity score (OR = 1.15, p = 0.04) and fibrinogen values (OR = 3.15, p = 0.04) at hospital admission.	Long Coronavirus disease
Bakilan et al., 2021³⁹	Turkey	Retrospective cross-sectional study	Eskisehir City Hospital	N = 280▫ Age Range: ≥ 18▫ Mean Age: 47.45 ± 13.92▫ Males: 97 (34.6%)▫ Females: 183 (65.4%)	At least 1 months after hospitalization	RT-PCR and/or SARS-CoV-2 antibody testing	● Subjective self-report during clinical follow-up○ Objective assessment via laboratory tests◘ Objective assessment of chest CT via CO-RADS	● Chest pain 30 (10.7%)● Arthralgia 101 (36.1%)● Muscle pain 143 (51.1%)● Spine pain 162 (57.1%)	Any premorbid disease condition, longer duration of hospitalization, increased D-dimer and/or decreased lymphocytes during acute phase of COVID-19 and positive COVID-19 findings in chest CT during acute COVID-19 infection were associated with pain symptoms onset or aggravation (Gender was not associated with difference in the reported rates of musculoskeletal pain symptoms)	Postacute COVID-19 patients
Carfi et al., 2020⁴⁰	Italy	Retrospective cohort study.	Fondazione Policlinico Universitario Agostino Gemelli IRCCS	N = 143▫ Age Range: ≥ 18▫ Mean Age: 56.5 ± 14.6▫ Males: 90 (62.9%)▫ Females: 53 (37.1%)	Individuals after discharge from hospital after recovery from COVID-19 and meeting WHO criteria for quarantine discontinuation^cMean days since:▫ Symptom onset: 60.3 ± 13.6▫ Discharge: 36.1 ± 12.9	RT-PCR	● Subjective assessment during clinical follow-up using a standardized questionnaire (of symptoms potentially correlated with COVID-19, nonvalidated)○ Objective assessment via EuroQol VAS (QoL)^d	● Joint pain 39 (27.3%)● Chest pain 32 (21.7%)● Headache 11 (7.1%)● Sore throat 9 (5.7%)● Myalgia 8 (5.3%)○ Worsened QoL observed in 63 (44.1%) patients		Persistent symptoms after recovery from acute COVID-19
Catalán et al., 2022⁴¹	Spain	Observational cohort study	Castellón General University Hospital	N = 76▫ Age Range: ≥ 18▫ Median Age: 61.5 (52.7–72.5)▫ Males: 33 (52.3%)▫ Females: 43 (47.7%)	> 1 year after admission for COVID-19.	RT-PCR	● Subjective self-report during clinical follow-up○ Objective assessment via SF-36 (QoL)	● Arthromyalgia 29 (38%)● Headache 13 (17.1%)● Arthritis 12 (16%)● Dysphagia 6 (7.89%)● Chest pain 5 (6.57%)● Odynophagia 2 (2.6%)○ Higher median scores in 4/8 SF-36 domains in steroid group steroids group in 4/8 SF-36 domains, after 1 year follow-up, of which only bodily pain and mental health domains showed statistical significance while general health and energy did not	Steroid treatment during acute phase of COVID-19 resulted in significantly less proportions of headache, dysphagia and chest pain at 1 year follow-up, as well as leading to attenuation of other persistent Long-COVID symptoms and improvement in QoL in the long term (compared to those not treated with steroids)	Long-COVID, postacute COVID-19 syndrome, Long-term persistence of COVID-19 symptoms
Daher et al., 2020⁴²	Germany	Case series	University Hospital RWTH Aachen	N = 33▫ Age Range ≥ 18▫ Mean Age: 64 ± 3 (qualifying into above 60 years group)▫ Males: 22 (67%)▫ Females: 43 (33%)	At least 6 weeks after hospital discharge.	RT-PCR	● Subjective self-report during clinical follow-up○ Objective assessment of functional outcomes and QoL impairments via the 6MWT, EQ-5D-5L, GAD-7, PHQ-9, SGRQ and of inflammatory markers via laboratory tests	● Sore throat 3 (9%)● Angina pectoris 6 (18%)● Myalgia 5 (15%)● Headache 5 (15%)● Stomach pains 1 (3%)○ 6MWT—26 (79%) under age-adjusted predicted values○ PHQ-9—patients suffered from mild depression [median = 7 (4–11)]○ GAD-7—patients mostly suffered from mild anxiety [median = 4(1–9)]○ SGRQ showed mainly reduced mobility domain [median = 54 (19–78)]○ EQ-5D-5L—slight to moderate problems with mobility, self-care (washing or dressing), doing usual activities, pain/discomfort and anxiety/depression○ Laboratory—At the time of follow-up ferritin, CRP, and IL-6 declined for most patients to normal values. Some had elevated D-dimer		Pulmonary and extrapulmonary sequelae of COVID-19
Fernández-de-Las-Peñas et al., 2021³³	Spain	Case–control study	Urban hospital in Madrid	N_total = 738 (total cohort)N_cases = 369▫ Age Range ≥ 18▫ Mean Age: 60 ± 15.5▫ Males: 176 (48%)▫ Females: 193 (52%)N_controls = 369▫ Age Range ≥ 18▫ Mean Age: 60 ± 15.0▫ Males: 176 (48%)▫ Females: 193 (52%)	> 7 months after hospital discharge	RT-PCR	● Subjective self-report via telephonic interview (using predefined list of general post-COVID symptoms and any other symptom being reported by the patient)	(For N_total)● Migraine-like headache & Tension type like headaches 84 (11.38%)● Musculoskeletal pain 210 (28.45%)● Thorax-Chest pain 31 (4.2%)● Widespread pain 43 (5.82%)	Myalgia during acute phase of COVID-19 was associated with preexisting history of musculoskeletal pain (half of which experienced exacerbation) and to significantly higher proportion of PCS musculoskeletal pain symptoms (compared to those without myalgia)	Long-term post-COVID sequelae, post-COVID symptoms
Fernández-de-Las-Peñas et al., 2021⁴³	Spain	Case–control study	Hospital Clínico San Carlos	N_total = 615 (total cohort)N_cases = 205▫ Age Range ≥ 18▫ Mean Age: 55.5 ± 14.0▫ Males: 82 (40%)▫ Females: 123 (60%)N_controls = 410▫ Age Range ≥ 18▫ Mean Age: 55.4 ± 14.0▫ Males: 164 (40%)▫ Females: 246 (60%)	> 7 months after hospital discharge	RT-PCR	● Subjective self-report via clinical follow-up (nonvalidated instrument but in systematic procedure)	(For N_total)● Tension-Type Like Headache 91 (14.7%)● Migraine-like Headache 15 (2.4%)	Headache at acute COVID onset was associated with increased rates of post-COVID symptoms (including new onset tension-type like headache and fatigue)	Post-COVID symptoms
Fernández-de-Las-Peñas et al., 2021⁴⁴	Spain	Multicenter case–control study	Three public hospitals in Madrid	N_total = 435 (total cohort)N_cases = 145▫ Age Range ≥ 18▫ Mean Age: 70.2 ± 13.2▫ Males: 90 (62.1%)▫ Females: 55 (37.9%)N_controls = 290▫ Age Range ≥ 18▫ Mean Age: ±▫ Males: 180 (62.1%)▫ Females: 110 (37.9%)	> 12 weeks after hospital discharge.	RT-PCR	● Subjective self-report via nonstandardized questionnaire during telephone interview○ Objective assessment via several items from the FIC (for functional outcomes)^e	(For N_total)● Musculoskeletal pain 192 (44.13%)● Migraine-like headache 9 (2.1%)Following number of patients had limitation with following activities○ 196 (45%) with at least one functional limitation ADLs.○ 64 (14.7%) with occupational activities.○ 159 (36.5%) with social/leisure activities.○ 100 (23%) with instrumental ADLs.○ 132 (19.4%) with basic ADLs.	No between-group (diabetic vs. nondiabetic) differences existed for all functional outcomes and post-COVID symptoms.Diabetes was not a risk factor for experiencing long-term post-COVID symptoms	Long-term Post-COVID Symptoms
González-Andrade, 2021⁴⁵	Ecuador	Observational, and cross-sectional study	Quito City	N = 1366▫ Age Range: 12–85▫ Mean Age: 39 ± 10▫ Males: 679 (49.71%)▫ Females: 687 (50.29%)	> 4 weeks after COVID-19 infection.	RT-PCR	● Subjective self-report via patients records or clinical follow-up○ Objective assessment via CFQ-11^f and subjective questionnaire and perception scale for physical and mental fatigue assessment^g	● Body pain 578 (42.31%)● Chest pain 209 (15.3%)● Headache 618 (45.27%)● Abdominal pain 156 (11.42%)● Myalgia 435 (31.84%)● Arthralgia 317 (23.22%)● Throat pain 262 (19.18%)○ Mean alteration of 6.8% (ranging from 0 to 21%) in ADLs was demonstrated in long-COVID patients.○ Impact on walking was reported in 20.7% of cases while for physical activities in 7.2% of cases	Sedentary lifestyle (walking < 30 min a day), being a health worker, age ≥ 55, HTN and CKD increased the risk for COVID-19 symptom persistence	Post-COVID-19 conditions
Graham et al., 2021⁴⁶	United States of America (Chicago)	Prospective study	Northwestern Memorial Hospital	N = 50 (SARS-CoV-2+ group):▫ Age Range: ≥ 18▫ Mean Age: 43.7 ± 11.8▫ Males: 17 (34%)▫ Females: 33 (66%)	At least 6 weeks from symptom onset	RT-PCR	● Subjective self-report during clinical follow-up.○ Objective assessment via PROMIS (for QoL) and Laboratory tests (inflammatory markers)	● Chest pain 14 (28%)● Headache 32 (64%)● Myalgia 30 (60%)● Pain other than chest 20 (40%)○ Impaired QoL in cognitive and fatigue domains was reported.○ Measured inflammatory markers were within the norm (CRP, ESR, Di-dimer, Ferritin)	“Long-haulers” cohort demonstrated increased F:M ratio (2.3:1), prevalence of premorbid autoimmune diseases and increased ANA titer compared to the general population, implying about autoimmune contribution.○ Premorbid depression/anxiety were more prevalent than in general population, implying possible neuropsychiatric vulnerability to becoming a “long hauler” after SARS-CoV-2 infection	Post-COVID syndrome, nonhospitalized long haulers
Hossain et al., 2021⁴⁷	Bangladesh	Prospective Cohort Study	24 testing facilities across Bangladesh	N = 2198▫ Age Range: ≥ 18▫ Mean Age: 38.4 ± 11.4▫ Males: 1591 (72.5%)▫ Females: 607 (27.5%)	31 weeks after positive diagnosis	RT-PCR	● Subjective self-report via non standardized questioner.○ Objective assessment using the PCFS (functional outcomes)^h and clinical parameters (of acute phase disease severity).◘ Objective laboratory data (blood group)	● Chest pain 1 (0.045%)● Headache 8 (0.36%)● Musculoskeletal (muscle) pain 60 (2.72%)○ Reported functional limitations in those with long COVID:73.3% (n = 261) had none.20.2% (n = 72) had negligible limitations.6.5% (n = 23) had slight limitations.○ Mean PCFS in long-COVID symptoms was 12.14 ± 8.8 (0 to 100 score range)	Risk factors for long-COVID symptoms were female gender, severeness of acute phase COVID-19, at least one chronic comorbidity, rhesus positive blood group factor, prior functional limitations and occupations such as frontline healthcare professional, police worker, housewife or private sector job holderRisk factors of longer duration of long COVID symptoms were young age, female gender, severity of acute phase COVID-19, current smoking, rural geographical residence and prior functional limitationsHigher mean PCFS scores for long COVID patients in this cohort indicating that functional limitations progress over time in people living with and affected by long COVID	Postacute COVID-19 symptoms, Long-COVID symptoms
Huang et al., 2021⁴⁸	China	Ambidirectional cohort study	Jin Yin-tan Hospital	N = 1272 (at 12 month)▫ Age Range ≥ 18▫ Median Age: 59 (49-67)▫ Males: 681 (53%)▫ Females: 591 (47%)	12 months after symptom onset	RT-PCR	● Subjective self-report during clinical follow-up (using nonstandardized symptom questionnaires)○ Objective assessment using EQ-5D-5L (for QoL), mMRC Dyspnea scale (functional disability), and 6MWD test (for functional capacity)	● Joint pain 157 (12%)● Chest pain 92 (7%)● Sore throat or difficult to swallow 44 (3%)● Myalgia 54 (4%)● Headache 61 (5%)○ EQ-5D-5L: COVID-19 survivors at 12 months had (compared to controls):Lower self-assessment on all scores of QoL than did controls.Increased risk for problems with mobility and painMore prevalent overall symptoms.(p < 0.0001)○ 6MWD of less than normal lower limit 12% (147/1248) 12-month post-COVID-19○ mMRC score of 1 or more with dyspnea in 30% (380/1271) 12-month post-COVID infection	Women had increased odds for persistence of fatigue, muscle weakness, anxiety, depression, as well as pulmonary diffusion impairment compared to malesIncreased age was positively associated with anxiety, depression and dyspnea with diffusion impairmentCorticosteroids therapy at acute phase was associated with increased risk of fatigue or muscle weakness	Hospital survivors with COVID-19
Karaarslan et al., 2021⁴⁹	Turkey	Single-center cohort study	Gulhane Training and Research Hospital	N = 300▫ Age Range: ≥ 18▫ Mean Age: 52.58 ± 12.01▫ Males: 179 (59.7%)▫ Females: 121 (40.3%)	1 month after hospitalization.	RT-PCR	● Subjective assessment via phone survey, using standardized form (severity of each COVID-19 symptom was queried 5 level Likert scale)	● Arthralgia, 66 (22.0%)● Headache 26 (8.7%)● Myalgia 63 (21.0%)● Low back pain 49 (16.3%)● Back pain 68 (22.7%)● Neck pain 31 (10.3%)● Sore throat 9 (3.0%)	Higher BMI was associated with increased odds for persistence of myalgia and arthralgia at 1-month post hospitalization	Post discharge symptomsPostacute Sequelae of SARS-CoV-2 infection (PASC)
Kayaaslan et al., 2021³⁴	Turkey	Prospective study	Ankara City Hospital	N = 1007▫ Age Range: ≥ 18▫ Mean Age: 45.0 ± 16.4▫ Males: 548 (54.4%)▫ Females: 459 (45.6%)	> 12 weeks from diagnosis	RT-PCR	● Subjective self-report via clinical follow-up or telephone (using nonstandardized questionnaire designed by the authors)○ Objective assessment using the C19-YRS scaleⁱ (Functional outcomes)	● Abdominal pain 4 (0.4%)● Chest pain 58 (5.8%)● Headache 57 (5.7%)● Myalgia 133 (13.1%)○ Patients with persistent symptoms reported mild-to-moderate level of impact on their level of function	Severe acute phase COVID-19, hospitalization, and presence of one or more medical comorbidities were independent risk factors for the development of persistent symptoms	Post-COVID syndrome, Long-COVID
Maamar et al., 2022⁵⁰	Spain	Cross-sectional study	Primary Health Care center (Camargo-Interior) in Santander	N = 121▫ Age Range: 18-88▫ Mean Age: 45.7 ± 16▫ Males: 53 (43.8%)▫ Females: 68 (56.2%)	> 3 months after the acute phase onset	RT-PCR	● Subjective self-report via interview (using nonstandardized symptom questionnaire).○ Objective assessment using laboratory tests (for inflammatory biomarkers: CRP, ferritin, LDH, D-dimer, fibrinogen, neutrophil and lymphocyte counts, NLR).○ Five different LGI composite indices C1-5^j assessed at least 3 months postacute phase resolution	● Odynophagia 12 (9.9%)● Myalgia 23 (19%)● Headache 21 (17.4%)● Chest pain 4 (3.3%)Women (N = 68)● Odynophagia 8 (11.7%)● Myalgia 16 (23.5%)● Headache 16 (23.5%)● Chest pain 4 (5.88%)Men (N = 53)● Odynophagia 4 (7.54%)● Myalgia 7 (13.2%)● Headache 5 (9.43%)○ LGI range CRP levels correlated with PCS in men (coefficient of 0.386, p = 0.011).Fibrinogen levels were higher in subjects with persistent myalgia (510 ± 82 mg/dL vs. 394.6 ± 87; p = 0.013)	Neutrophil counts, NLR, CRP and fibrinogen showed the best correlations with PCS symptom persistenceAll 5 LGI composite indices (C1–5) significant correlation with PCS (C1,3,4 more frequently met in women while C2,5 in males)Symptoms and LGI biomarkers of PCS demonstrate gender differences	Post-COVID syndrome
Mendelsohn et al., 2022⁵¹	South Africa	Retrospective cross-sectional study	Retreat Community Health Centre (CHC)	N = 174▫ Age Range: ≥ 18▫ Mean Age: 50.3 ± 13.6▫ Males: 66 (37.9%)▫ Females: 108 (62.1%)	2 months postdiagnosis	RT-PCR	● Subjective assessment via standardized survey (nonvalidated).○ Objective assessment via mMRC and WHO performance status classification scale (overall activity level assessment)^k	● Body aches 26 (14.9%)● Chest pain 19 (10.9%)● Headache 27 (15.5%)○ 34.5% (60/174) reported fatigue during follow-up with mean reported activity level score of 1 (0.7) on the WHO/ECOG scale (ambulatory and able to carry out sedentary or light work while restricted in physically strenuous activity)○ Mean grade 1 (1.0) on the mMRC reported by those with dyspnea symptoms (35/174, 20.1%)		Long-COVID symptoms
Naik et al., 2021⁵²	India	Prospective observational study	Tertiary care center in Delhi	N = 1234▫ Age Range: ≥ 18▫ Mean Age: 41.6 ± 14.2▫ Males: 856 (69.4%)▫ Females: 378 (30.6%)	At least 2 weeks post laboratory-confirmed diagnosis(median follow-up days 91 (45–181))	RT-PCR	● Subjective self-report during clinical or telephone follow-up (using non standardized questionnaires)	● Chest pain 6 (0.5%)● Myalgias 54 (4.4%)	Hospitalization, and hypoxia (as a measure for disease severity) or hypothyroidism during the acute episode were identified as significant risk for long COVID	Post-COVID-19 sequelae, post-COVID-19 syndrome, Long COVID Syndrome
Nehme et al., 2021⁵³	Switzerland	Single-center cohort study	Geneva University Hospitals	N_total = 629 (Total cohort):▫ Age Range ≥ 18▫ Mean Age: 42.1 ± 13.5▫ Males: 246 (39.1%)▫ Females: 383 (60.9%)N_{7–9 month} = 410 (7–9 month cohort)	7–9-month after diagnosis of symptomatic COVID-19	RT-PCR	● Subjective self-report via semistructured phone follow-up, using standardized interview (nonvalidated; Likert scale to grade intensity of each symptom, except back pain).○ Objective assessment using the ECOG performance scale^l and mMRC dyspnea scale (for functional outcomes)	(For N_{7–9 month})● Headache 41 (10%)● Myalgia 40 (6.3%)● Arthralgia 14 (3.4%)● Chest pain 14 (3.2%)● Back Pain 11 (2.7%)● Throat Pain 10 (2.4%)○ 20.7% (85/410) reported some degree of fatigue27% (23/85) of those reported fatigue symptoms, graded the ECOG as 1 (restricting strenuous activity)○ 11.7% of patients (48/410) reported at least some degree of dyspneaOf these, 34 reported dyspnea with a score of 1 or higher on the mMRC scale (8.29% of total patients 34/410)		Long-COVID, postacute sequelae of SARS-CoV-2 infection, or post-COVID-19 condition
Och et al., 2021⁵⁴	Poland	Longitudinal Cohort Study	7th Naval Hospital in Gdansk	N_total = 79 (Total cohort):▫ Age Range: ≥ 18▫ Median Age: 70 (64–76.5)▫ Males: 39 (49.37%)▫ Females: 40 (50.63%)N_{6 month} = 73 (6 month follow-up cohort)	3 months and 6 months after diagnosis of COVID-19	RT-PCR	● Objectiveself-report via phone interview using the SRSQ^m○ Objective assessment using mMRC dyspnea scale, EQ-5D-5L and EQ-VAS	(For N_{6 month})● Myalgia 15 (20.55%)● Headache 17 (23.29%)● Joint pain 16 (21.92%)● Chest pain 13 (17.81%)● Sore throat or difficult to swallow 6 (8.22%)○ 34.25% had mMRC scale grade of at least 1 with dyspnea (compared to 21.5% before COVID-19)○ Decrease in QoL reported for all domains of EQ-5D-5L, mainly in pain/discomfort and anxiety dimensions○ Mean EQ-VAS score was 64.4 ± 16.2 (vs. 69.9 ± 17.6 prior to infection; p < 0.001)		Post-COVID-19 Syndrome
Osikomaiya et al., 2021¹¹	Nigeria	Retrospective study	Mainland Hospital Yaba	N = 274▫ Age Range: 10-83▫ Mean Age: 41.8 ± 11.8▫ Males: 181 (66.1%)▫ Females: 93 (33.9%)	At least 2 weeks after discharge from COVID-19 isolation (by meeting prespecified criteria)ⁿ	RT-PCR	● Subjectiveself-report during clinical follow-up	● Chest pain 27 (9.8%)● Headache 35 (12.8%)● Sore throat 11 (4.0%)● Myalgia 24 (8.8%)	More symptomatic patients with moderate or severe forms of COVID-19 disease (based on WHO criteria¹¹ at initial diagnosis had higher odds for persistent COVID-19 like symptoms post discharge (compared to the mildly symptomatic ones)	Persistent COVID-19, Persistent COVID-19 like symptoms
Romero-Duarte et al., 2021⁵⁵	Spain	Retrospective observational study	Hospital San Cecilio, Reina Sofía, Jaén and Puerto Real	N = 797▫ Age Range ≥ 18▫ Mean Age: 63▫ Males: 428 (53.7%)▫ Females: 369 (46.3%)	6 months following hospital discharge	RT-PCR	● Subjective(Sequelae or persistent symptomatology (SPS) gathered from primary care records)	● Musculoskeletal pain 122 (15.3%)● Rib pain 36 (4.5%)● Thoracic pain 53 (6.6%)● Headache 42 (5.3%)● Abdominal pain 43 (5.4%)	Women presented higher frequencies of headache.Requiring hospitalization during acute phase was associated with high frequency of at least one persistent symptom (63.9%)Persistence of thoracic pain was one of the main factors to be independently associated with increased risk of returning to emergency services post the acute phase (thus may serve as indicator of potential severity of PCS)	Sequelae or persistent symptomatology (SPS), Long-term COVID, Long-term effects of COVID
Taha et al., 2021⁵⁶	Egypt	Cross-sectional	Ain Shams University Hospital	N = 100▫ Age Range: ≥ 18▫ Mean Age: 57 ± 17.74▫ Males: 61 (61%)▫ Females: 39 (39%)	6 months post-COVID recovery	RT-PCR	● Subjective assessment via self-report during clinical follow-up○ Objective assessment using laboratory tests (for inflammatory biomarkers)	● Myalgia 24 (24%)● Arthralgia 62 (62%)	Older age, arthralgia, smoking status, elevated IL-6 during acute infection, elevated ESR and CRP 6 months postacute phase COVID-19 were significantly associated with occurrence of post-COVID-19 arthritis	Post-COVID
Taquet et al., 2021²⁸	United States of America	Retrospective cohort study	TriNetX Analytics	N = 273,618▫ Age Range: ≥ 10▫ Mean Age: 46.3 ± 19.8▫ Males: 121,461 (44.4%)▫ Females: 152,157 (55.6%)	3-6 months after COVID-19 diagnosis	Electronic health records analysis showing ICD-10 code of confirmed COVID-19 diagnosis	● Subjective assessment usingprimary care electronic health records	● Chest/Throat pain15,624 (5.71%)● Myalgia 4214 (1.54%)● Headache 12,669 (4.63%)● Pain—19,674 (7.19%)	Significantly increased risk for overall long-COVID features (including pain symptoms) were associated with sex (females), increased age, and increased severity of acute COVID-19 illness.Women, young individuals and those with less severe disease (nonhospitalized or those without leukocytosis at acute phase) had increased risk for pain symptoms specifically	Long-COVID
Tetik et al., 2021⁵⁷	Turkey	Descriptive study.	Malatya Training and Research Hospital	N = 152▫ Age Range: ≥ 18▫ Mean Age: 44.9 ± 13▫ Males: 86 (56.6%)▫ Females: 66 (43.4%)	At least 1 month from recovery and hospital discharge	RT-PCR	● Subjective assessment using via telephone interview using symptom questionnaire (nonvalidated)	● Joint pain 1 (0.7%)● Headache 2 (1.3%)● Body pain 6 (3.9%)● Lumbar spine pain 4 (2.6%)	Old age was associated with increased rates of post-COVID pain symptoms (especially for those > 50 years compared to the rest, p = 0.025).Increased BMI and chronic disease(s) were associated with higher proportion of PCS pain symptom persistence, though nonsignificantly	Long-COVID, Post-COVID, Post-COVID pain, Prolonged COVID
Zayet et al., 2021²⁷	France	Observational retrospective study	Nord Franche-Comté Hospital	N = 354▫ Age Range: 19-98▫ Mean Age: 49.6 ± 18.7▫ Males: 131 (37%)▫ Females: 223 (63%)	At least 9 months after COVID-19 onset	RT-PCR	● Subjective assessment using online nonstandardized questionnaires○ Objective assessment using laboratory values (for inflammatory biomarkers, through medical records)	● Myalgia 33 (9.32%)● Headache 33 (9.32%)● Arthralgia 31 (8.75%)● Chest Pain 15 (4.23%)● Sore Throat 9 (2.54%)● Abdominal Pain 8 (2.26%)● Facial Headache 6 (1.69%)● Dysphagia 6 (1.69%)	Suffering from chronic rhinosinusitis, health care workers, hospitalization during acute phase of disease and/or had longer hospitalization period were associated with higher rates of persistent post-COVID symptoms	Post-COVID-19 Syndrome

Note: Unless otherwise specified and/or only when provided: median values reported as median (interquartile range); means values reported as mean ± standard deviation; Age is in years; When more than one cohort size is described in the sample characteristics column, the one used for meta-analysis is mentioned in the outcome column.
Abbreviations: 6MWT, 6-min walk test (Tests individuals for functional capacity; 6MWD: distance walked in 6 min); CO-RADS, COVID-19 Reporting and Data System; CRP, C-reactive protein; EQ-5D-5L, EuroQol-5 Dimension-5 levels; EuroQol VAS, EuroQol visual analog scale; IL-6, interleukin-6; LCS, long-COVID syndrome; LDH, lactate dehydrogenase; LGI, low-grade inflammation; mMRC scale, Modified Medical Research Council scale; NLR, neutrophil to lymphocyte ratio; PHQ-9, Patient Health Questionnaire 9; PROMIS, Patient-Reported Outcomes Measurement Information System; RT-PCR, Reverse transcription-polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SF-36, 36-Item Short Form Survey; SGRQ, St. George’s respiratory questionnaire.
^a Nonstandardized 5 item (questions) Likert scale questionnaire—Patients were asked to compare their current status (on five items) with their pre-COVID-19 status based on a Likert scale (1. Much worse; 2. Somewhat worse; 3. The same as before; 4; Somewhat better; 5. Much better). The 5 item were: 1. ability to perform the activity of daily living; 2. concentration and mind workability; 3. studying and reading ability; 4. QoL; 5. hope for the future.
^b CT severity scores were calculated using the method described by Pan et al. CT: Computed tomography. Scores were defined for each lobe of the lung and the sum of the scores of the lobes constitute the total lung score. Total score scale: 0–25. Increased score = increased severity.⁵⁸
^c All patients who met World Health Organization criteria for discontinuation of quarantine (no fever for 3 consecutive days, improvement in other symptoms, and 2 negative test results for severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2] 24 h apart) were followed up.
^d EuroQol VAS—EuroQol visual analog scale—Worsened QoL was defined by a 10-point difference in health status before COVID-19 vs. at the time of the visit.
^e Functional Impairment Checklist (FIC)—disease-specific tool for evaluating the functional consequences of SARS (15): the items that were used were dyspnea at rest, dyspnea on exertion, generalized fatigue, limitations occupational activities, limitations social/leisure activities, limitations in basic activities of daily living, and limitations in instrumental activities of daily living.
^f Chalder Fatigue Scale (CFQ-11) to assess fatigue: 33 points indicated severe fatigue; 22–33 points, moderate fatigue; below 22 points, no fatigue.
^g Subjective perception scale for fatigue assessment—patient asked to indicate the average intensity of your fatigue in the last 24 h on a scale of 0 to 10. Where: 0 = no fatigue and 10 = maximum fatigue you can imagine.
^h Post-COVID-19 Functional Status Scale (PCFS) (Measurement of cardiorespiratory parameters for determining impact on function and disability in by covering 6 domains: survival; constant care; basic activities of daily living; instrumental activities of daily living; participation in usual social roles and a symptom checklist).
ⁱ The COVID-19 Yorkshire Rehabilitation Scale (C19-YRS)—22-item patient-reported outcome measure to evaluate long-term impact of COVID-19 across various domains of Activities and Participation of the International Classification of Functioning, Disability, and Health and evaluate the impact of PCS rehabilitation.⁵⁹
^j LGI composite indices C1-5 (based on combinations between NLR, Neutrophil count, CRP and fibrinogen levels) assessed at least 3 months postacute phase resolution.
^k WHO/ECOG performance status scale [homepage on the Internet]. [cited 2021 Sep 12]. Available from: https://ecog-acrin.org/resources/ecog-performance-status.
^l Eastern Cooperative Oncology Group (ECOG) performance scale—describes the degree of functional status in terms of daily activities and physical performance.
^m Self-reported symptoms questionnaire (SRSQ)—according to Huang et al.
ⁿ Discharge from COVID-19 isolation facilities between March and May 2020 was based on the World Health Organization (WHO) criteria: Three days after resolution of symptoms and/or Two negative RT-PCR SARS-CoV-2 results, at least 24 h apart.

Quality of evidence and risk of bias assessment

The methodological quality and risk of bias of the included studies were evaluated using the Newcastle-Ottawa Scale (NOS),⁶⁰ modified for case–control, cohort, and cross-sectional studies. All other studies, including those with unclear study designs, were assessed according to the prospective cohort NOS. Two reviewers (OK and EB) rated the studies independently, corroborated the results, and resolved discrepancies through discussion to reach mutually agreed upon ratings (Appendix S1).

Data synthesis and analysis

The meta-analysis (MA) was carried out using R version 4.2.1 (R Core Team, 2022). Statistical significance was set at an α level of 0.05. Assuming the existence of substantial variability within and between component studies, and in order to facilitate the generalization of their results, a priori adoption of the random-effects-restricted maximum-likelihood (REML) model was made. The statistical model was used to compute the pooled proportions, providing us with the estimated mean values and distributions of the true effect sizes.^61, 62 The meta::metaprop function⁶³ was used to calculate pooled proportions via the REML model on logit-transformed values of the single proportions from each study, which were subsequently transformed back into proportional variables. Logit transformation was performed to overcome possible skewing by proportional values close to zero or one.^64, 65 Clopper-Pearson & Wald-95% confidence intervals (CIs) were calculated for individual studies and pooled proportions, respectively. Forest plots for our primary outcomes were generated through the meta::forest function.

Pain symptoms were stratified into six groups of major domains (our co-primary outcomes) for the reported pain symptoms. These allocations were decided and agreed upon after mutual discussion based on the preliminary literature search, which indicated that most reported pain symptoms could be grouped under either one of the broader categories provided by the major domains of joint pain, muscle pain, nervous system pain, gastrointestinal pain, cardiovascular & respiratory (under chest pain) systems, and general body pain (GBP). It was decided that if a reported pain symptom will not be amenable for inclusion in any one of the groups, it will be reported and, if necessary, addressed narratively. To avoid data duplication and effect size skewing, if study reported two or more pain symptoms under the same domain without providing sufficient information for identifying the exact co-occurrence of these symptoms per each individual, then only the largest group was included in analyses. When more than one follow-up period was provided for the same pain symptom, then the longest one since the infection was selected. Cases where symptoms were reported to be present before the COVID-19 acute phase without any sequential exacerbation were not included. The terms used to describe pain symptoms and the definitions for each major pain domain, as well as their mode of ascertainment (MoA), are reported in Table 1. A short description of the secondary outcomes is also provided, though, due to the very limited amount of data, it was decided to describe these narratively, as previously mentioned (Table 1).

In the subgroup analysis, the inverse variance method, the restricted maximum-likelihood estimator for τ², the Q-profile method for the confidence interval of τ² and τ, and the Hartung-Knapp adjustment for the random-effects model were applied as an update on the already performed meta-analyses.⁶⁶ Subgroups were defined by stratifying for sex, age, follow-up duration, MoA, hospitalization status, NOS rating category, and study design and conducted under the byvar argument, assuming distinct estimates of between-study variance for each subgroup. Studies were allocated to the 18–60 age group if no clear distinction was made between those above 60 and below 18 years of age. Studies that reported hospitalized and nonhospitalized cases, without clearly stratifying them were excluded. NOS ratings and study design subgroup analysis were used for sensitivity analysis. Anticipating high levels of disagreement between the independent raters in the individual scorings, the NOS category strata were decided to be divided into high scores and to moderate and low scores.^67–69 We assumed that under these circumstances high scored studies might better separate themselves from the rest, which, in turn, will facilitate some heterogeneity reduction. The statistical significance of the difference between effect sizes of the subgroups within each strata was calculated by using the Wald-type χ² test.

Moderate (30%), substantial (50%), and considerable (70%) cutoff values for heterogeneity (I²) were determined as recommended by GRADE criteria and the Cochrane handbook for heterogeneity interpretation.^70, 71 Visual assessment of funnel plots (generated by meta:metafunnel function) asymmetry was used for publication bias evolution, due to the inherently intuitive interpretation provided by this method compared to others which are usually more accurate.^72–74

RESULTS

Study selection and search results

611 abstracts were obtained from the initial search. After duplicate removal and by abstract and title screening and exclusion, 79 articles were fully read and assessed for eligibility. Twenty-six studies met full inclusion criteria (Figure 1) and included 10 prospective studies, six cross-sectional studies, six retrospective cohort studies, three case controls, and one case series. These were grouped into four study design groups (Appendix S1). Component studies were used to illustrate through quantitative synthesis the effect of COVID-19 infection on pain outcomes. Qualitative analysis via narrative description was done for secondary outcomes (eg, QoL, functional impairments, and inflammatory markers).

Details are in the caption following the image — **FIGURE 1**Open in figure viewer PowerPointA flow diagram of the article selection process from the bibliographic search of electronic databases. Reported in line with PRISMA guidelines.

Study characteristics

Six studies analyzed data from Spain, five from Turkey, two from the US, and one from Italy, Switzerland, France, Nigeria, China, Iran, India, South Africa, Egypt, Ecuador, Bangladesh, Germany, and Poland. Sample sizes ranged from 33 to 273,618, and follow-up periods were reported for up to 12 months. All studies reported regarding the MoA of COVID-19 infection. More detailed description is presented in Table 2.

Quality of evidence and risk of bias assessment

The Overall NOS ratings for the included studies were moderate to low, reflected by the mean scores of 5.72 out of 9, 6.3 out of 8, 3 out of 6, and 6.66 out of 9 for prospective cohorts, retrospective case–control, retrospective cohort, and cross-sectional study design groups, respectively. Notable methodological pitfalls demonstrated most often by the studies were lack of nonexposed cohort inclusion, lack of adequate representation of target population by the selected cohort (eg, only hospitalized), not providing an explanation as to the sample size selection, and not mentioning whether symptoms were present before acute infection and utilizing nonstandardized methods for our primary outcome assessment. For the secondary outcomes, however, most component studies used validated assessment methods. Scores for all component studies are included in Appendix S1. Stratifying by NOS categories did not demonstrate any significant difference in the sensitivity analysis across all major pain domains (Table 3).TABLE 3. Subgroup and sensitivity analyses for the primary outcomes.

	No. of studies	Proportion	95% CI	I² (%)	Q	p Subgroup (χ² test)
Chest pain
Sex
Females	3	0.0834	0.0167; 0.3270	90.3	30.96	0.9315
Males	3	0.0582	0.0013; 0.7428	91.4	35.00
Age group
Below 18 years	–	–	–	–	–	0.0850
18–60 years	20	0.0846	0.0537; 0.1309	97.5	866.94
60 or above years	1	0.1818	0.0840; 0.3499	–	–
COVID-19 hospitalization status
Hospitalized	11	0.0787	0.0441; 0.1367	94.5	199.58	0.9053
Nonhospitalized	5	0.0908	0.0297; 0.2458	92.4	52.65
Follow-up duration
< 6 Months	14	0.0877	0.0445; 0.1656	97.5	601.83	0.9462
≥ 6 Months	7	0.0856	0.0570; 0.1268	92.3	91.23
Mode of ascertainment
Subjective	17	0.0860	0.0528; 0.1371	97.5	804.68	0.8572
Objective	4	0.0945	0.0110; 0.4946	95.0	39.71
NOS rating category
High	8	0.0698	0.0405; 0.1176	90.9	121.44	0.2564
Moderate and Low	13	0.1096	0.0539; 0.2099	92.5	147.11
Study design
Prospective Cohort	9	0.0499	0.0160; 0.1446	92.7	110.34	< 0.0001
Retrospective Cohort	6	0.0771	0.0554; 0.1064	92.2	51.41
Cross-sectional	5	0.1310	0.0702; 0.2316	84.0	61.87
Retrospective Case–control	1	0.1809	0.0229; 0.6754	–	–
Gastrointestinal system pain
Sex
Females	3	0.0573	0.0152; 0.1931	75.9	8.31	0.5957
Males	3	0.0422	0.0183; 0.0945	46.0	3.71
Age group
Below 18 years	–	–	–	–	–	0.1375
18–60 years	13	0.0526	0.0280; 0.0967	98.3	725.40
60 or above years	1	0.1212	0.0462; 0.2818	–	–
COVID-19 hospitalization status
Hospitalized	9	0.0439	0.0336; 0.0572	55.2	17.85	0.4574
Nonhospitalized	3	0.0969	0.0054; 0.6794	97.9	95.22
Follow-up duration
< 6 Months	8	0.0417	0.0158; 0.1055	85.2	40.48	0.3108
≥ 6 Months	6	0.0710	0.0294; 0.1615	99.0	608.13
Mode of ascertainment
Subjective	10	0.0624	0.0298; 0.1262	98.5	656.73	0.1185
Objective	4	0.0335	0.0126; 0.0863	58.1	4.77
NOS rating category
High	5	0.0409	0.0195; 0.0838	90.6	84.92	0.0928
Moderate + Low	9	0.0957	0.0312; 0.2581	96.2	106.55
Study design
Prospective Cohort	7	0.0296	0.0101; 0.0839	83.9	31.03	0.0449
Retrospective Cohort	5	0.0615	0.0213; 0.1652	93.5	46.38
Cross-sectional	2	0.1875	0.0004; 0.9925	83.9	31.03
Retrospective Case–control	–	–	–	–	–
Musculoskeletal system pain—Joint related
Sex
Females	2	0.1453	0.0090; 0.7609	94.4	35.53	0.7647
Males	2	0.1005	0.0029; 0.8093	93.1	29.04
Age Group
Below 18 years	–	–	–	–	–	–
18–60 years	12	0.1618	0.1012; 0.2485	94.2	223.02
60 or above years	–	–	–	–	–
COVID-19 hospitalization status
Hospitalized	8	0.1551	0.0735; 0.2981	91.0	77.73	0.3086
Nonhospitalized	2	0.0950	0.0046; 0.5327	98.3	58.98
Follow-up duration
< 6 Months	5	0.1998	0.1457; 0.2679	76.4	16.93	< 0.0001
≥ 6 Months	7	0.1211	0.0488; 0.2700	95.2	146.91
Mode of ascertainment
Subjective	8	0.1711	0.1033; 0.2699	94.0	166.68	0.6705
Objective	4	0.1349	0.0114; 0.6784	96.4	56.33
NOS rating category
High	5	0.1330	0.0687; 0.2417	90.0	49.84	0.4191
Moderate + Low	7	0.1839	0.0838; 0.3570	92.0	87.39
Study design
Prospective Cohort	6	0.0888	0.0273; 0.2528	92.1	63.17	0.0313
Retrospective Cohort	3	0.1805	0.0134; 0.7807	87.8	8.18
Cross-sectional	3	0.2480	0.1508; 0.3799	94.4	71.32
Retrospective case–control	–	–	–	–	–
Musculoskeletal system pain—muscle related
Sex
Females	4	0.1729	0.0966; 0.2903	74.9	11.94	0.1394
Males	4	0.1123	0.0606; 0.1988	62.2	7.93
Age group
Below 18 years	–	–	–	–	–	0.6676
18–60 years	22	0.1844	0.1175; 0.2773	99.8	11772.04
60 or above years	1	0.1515	0.0645; 0.3162	–	–
COVID-19 hospitalization status
Hospitalized	14	0.1367	0.0650; 0.2649	99.8	5827.05	0.2088
Nonhospitalized	4	0.2594	0.0680; 0.6269	97.0	98.79
Follow-up duration
< 6 Months	14	0.1598	0.0869; 0.2753	99.7	5124.76	< 0.0001
≥ 6 Months	9	0.1929	0.0900; 0.3659	98.7	681.22
Mode of ascertainment
Subjective	20	0.1859	0.1173; 0.2817	99.8	10831.45	0.8291
Objective	3	0.1608	0.0070; 0.8390	98.9	176.29
NOS rating category
High	9	0.1565	0.0843; 0.2723	99.8	4848.34	0.4246
Moderate + Low	14	0.2152	0.1086; 0.3817	98.0	555.88
Study design
Prospective Cohort	11	0.1541	0.0722; 0.2991	98.3	572.52	0.0216
Retrospective Cohort	6	0.0979	0.0281; 0.2897	99.8	2143.11
Cross-sectional	4	0.3025	0.2042; 0.4229	91.2	56.92
Retrospective Case–control	2	0.3091	0.0119; 0.9434	97.5	81.01
General body pain
Sex
Females	–	–	–	–	–	–
Males	–	–	–	–	–
Age group
Below 18 years	–	–	–	–	–	–
18–60 years	6	0.1701	0.0644; 0.3791	99.7	1770.74
60 or above years	–	–	–	–	–
COVID-19 hospitalization status
Hospitalized	3	0.0886	0.0147; 0.3878	96.2	52.43	0.0090
Nonhospitalized	3	0.3059	0.0787; 0.6947	95.3	42.36
Follow-up duration
< 6 Months	4	0.1659	0.0219; 0.6388	97.2	70.31	0.9702
≥ 6 Months	2	0.1704	0.0105; 0.7989	97.4	77.18
Mode of ascertainment
Subjective	5	0.1743	0.0487; 0.4653	99.8	1757.47	0.7392
Objective	1	0.1701	0.0644; 0.3791	–	–
NOS rating category
High	4	0.1786	0.0000; 0.9997	98.2	55.55	0.9246
Moderate + Low	2	0.1663	0.0375; 0.5050	97.3	111.93
Study design
Prospective Cohort	2	0.1408	0.0000; 1.0000	96.7	30.23	< 0.0001
Retrospective Cohort	1	0.0719	0.0709; 0.0729	–	–
Cross-sectional	2	0.2684	0.0006; 0.9958	97.6	42.36
Retrospective Case–control	1	0.2067	0.1571; 0.2671	–	–
Nervous system pain
Sex
Females	3	0.1505	0.0556; 0.3474	87.9	24.77	0.1231
Males	3	0.0648	0.0239; 0.1636	70.3	10.10
Age group
Below 18 years	–	–	–	–	–	0.5502
18–60 years	22	0.1151	0.0771; 0.1684	99.2	3707.95
60 or above years	1	0.1515	0.0645; 0.3162	–	–
COVID-19 hospitalization status
Hospitalized	15	0.0811	0.0553; 0.1176	97.2	493.69	0.0154
Nonhospitalized	5	0.2634	0.0901; 0.5637	98.0	203.84
Follow-up duration
< 6 Months	14	0.1304	0.0799; 0.2055	97.9	653.05	0.4530
≥ 6 Months	9	0.0982	0.0483; 0.1893	99.0	981.72
Mode of ascertainment
Subjective	18	0.1189	0.0751; 0.1832	99.4	3638.00	0.5457
Objective	5	0.1024	0.0695; 0.1483	56.8	6.95
NOS rating category
High	9	0.1020	0.0556; 0.1798	98.2	728.38	0.4039
Moderate + Low	14	0.1369	0.0823; 0.2190	97.8	507.28
Study design
Prospective Cohort	10	0.0956	0.0402; 0.2107	95.0	182.0	0.0382
Retrospective Cohort	6	0.1015	0.0409; 0.2304	98.7	312.61
Cross-sectional	4	0.2063	0.1244; 0.3224	97.9	290.25
Retrospective Case–control	3	0.0756	0.0212; 0.2362	94.2	51.35

p ≤ 0.05 indicating signficant difference between sub-groups (are bolded).

Results synthesis

Pooled proportions for the co-primary outcomes indicated that 8%, 6%, 18%, 18%, 17%, and 12% of subjects continued experiencing chest, gastrointestinal, joint, muscle, general body, and nervous system-related pain symptoms, respectively, up to 1 year after acute phase resolution of COVID-19. Elevated inflammatory biomarkers were reported for some of the individuals in five out of seven studies. Functional outcomes and/or QoL measures were examined in 13 studies, all reporting mild-to-marked impairments for various proportions of included cohorts.

Chest pain symptoms

The pooled proportion of individuals experiencing chest pain symptoms up to one year after COVID-19 acute phase resolution was 0.08 (95% CI, 0.06, 0.13; p < 0.01; n = 286,330; Figure 2). Only the study design strata demonstrated significant differences in the reported proportions, ranging between 0.181 and 0.049 for the various study designs (p_subgroup <0.0001; Table 3). Individuals above 60 years old demonstrated much higher pooled proportions than those in the 18–60 age group (0.181 and 0.084, respectively; p_subgroup = 0.085; Table 3).

Gastrointestinal system pain symptoms

The pooled proportion of individuals experiencing GI pain symptoms up to one year after COVID-19 acute phase resolution was 0.06 (95% CI, 0.03, 0.10; p < 0.01; n = 8222; Figure 3). Only the study design strata demonstrated significant differences between intersubgroup proportions, with ranging values between 0.187 and 0.029 for the various study designs (p_subgroup = 0.0449; Table 3). Those above 60 years demonstrated pooled rates of more than double of those in the 18–60 age group (0.121 and 0.052, respectively; p_subgroup = 0.137; Table 3).

Musculoskeletal system joint pain symptoms

The pooled proportion of individuals experiencing joint pain up to one year after COVID-19 acute phase resolution was 0.18 (95% CI, 0.11, 0.25; p < 0.01; n = 6522; Figure 4). Significantly higher proportions were reported during the first six months compared to the successive six-month period (0.199 and 0.121, respectively; p_subgroup <0.0001; Table 3). Study design strata also demonstrated significant intersubgroup differences, with pooled proportions ranging between 0.088 and 0.248 (p_subgroup = 0.031; Table 3).

Musculoskeletal system muscle pain symptoms

The pooled proportion of individuals experiencing muscle-type pain up to one year after COVID-19 acute phase resolution was 0.18 (95% CI, 0.13, 0.26; p < 0.01; n = 287,027; Figure 5). Much higher proportions were demonstrated for nonhospitalized individuals compared to hospitalized (0.259 and 0.136, respectively; p_subgroup = 0.208; Table 3). Follow-up duration and study design strata demonstrated significant intersubgroup differences for the pooled proportions (p_subgroup <0.0001 and p_subgroup = 0.026, respectively; Table 3). Lower proportions were reported during the first 6 months (0.159) compared to the following 6 months (0.192). Substantial range of pooled proportions (0.097–0.309) for reported muscle pain symptoms was demonstrated by the various study designs.

General body pain symptoms

The pooled proportion of individuals experiencing general body pain symptoms up to one year after COVID-19 acute phase resolution was 0.17 (95% CI, 0.07, 0.36; p < 0.01; n = 276,098; Figure 6). Stratifying by study design provided significant difference (p_subgroup <0.0001) for the intersubgroup pooled proportions, ranging from 0.071 to 0.268 (Table 3). Substantial and significant difference was demonstrated between the nonhospitalized and hospitalized strata pooled proportions (0.305 and 0.088, respectively; p_subgroup = 0.009; Table 3).

Nervous system pain symptoms

The pooled proportion of individuals experiencing nervous system pain (mostly migraine and tension-type headache) up to one year after COVID-19 acute phase resolution was 0.12 (95% CI, 0.08, 0.16; p < 0.01; n = 286,318; Figure 7). Females demonstrated higher proportion in reporting these symptoms compared to males (0.1505 and 0.0648, respectively; p_subgroup = 0.123; Table 3). Hospitalization status and study design strata demonstrated significant differences between intersubgroup proportions. Pooled proportions for the nonhospitalized were higher than those for the hospitalized (0.263 and 0.081, respectively; p_subgroup = 0.015; Table 3). As for the study design strata, range of pooled proportions was between 0.075 and 0.206 (p_subgroup = 0.0382; Table 3).

Heterogeneity

Substantial levels of heterogeneity were demonstrated across all meta-analyses for the pain domains, that is, chest pain (I² = 97%), GI pain (I² = 98%), joint-type pain (I² = 94%), muscle-type pain (I² = 100%), GBP (I² = 100%) and nervous system pain (I ² = 99%). Select strata within the following subgroup analyses demonstrated some degree of heterogeneity reduction, though these still remained in the moderate to considerable range (Table 3).

Publication bias

Funnel plots for all major pain domain meta-analyses demonstrated considerable asymmetry (Figures S1–S6).

Functional outcomes and quality of life

Functional impairments and/or QoL measures were reported in 13 studies, with some implementing more than one assessment instrument. All studies assessed these secondary outcomes for long-COVID in general reported mild-to-marked impairments of various proportions (Table 2).

One study used the 36-Item Short Form Survey (SF-36) to demonstrate that steroidal intervention at acute COVID-19 episodes translates into improved QoL outcomes during one-year follow-up, compared to those not treated.⁴¹ Three studies used the European QoL 5-dimension 5-levels (EQ-5D-5L) scale to report mild-to-marked problems in QoL for long-COVID patients,^42, 48, 54 two of which reported a decrease across all QoL scores by all long-COVID cohort.^48, 54 Two studies used the European QoL visual analog scale (EQ-VAS). one study reported worsened QoL in 44.1% of long-COVID individuals,⁴⁰ while the second study reported significantly lower mean EQ-VAS scores at six-month follow-up compared to those before acute COVID-19 infection.⁵⁴ One study used the Patient-Reported Outcomes Measurement Information System (PROMIS) to evaluate overall QoL, demonstrating notable impairments in the cognitive and fatigue domains.⁴⁶ The St. George’s respiratory questionnaire (SGRQ) was used in one study, showing that decreased mobility (mean activity score of 54) was the main contributor to overall reduced QoL experienced by long-COVID individuals.⁴² The 6 min walking test (6MWT) was used in two studies, which reported that 79% and 12% of acute COVID-19 survivors scored under their age-adjusted predicted values after six weeks and one-year follow-up, respectively^42, 48 Four studies used the modified Medical Research Council (mMRC) dyspnea scale to assess the degree of baseline functional disability attributed to breathlessness. A mean score of ≥ 1 was reported by 11.7%–34.25% of acute COVID-19 survivors during two-month to one-year follow-up periods.^{48, 51, 53, 54} The Eastern Cooperative Oncology Group (ECOG) performance scale was used by two studies, which reported that 20.7-34.5% of acute COVID-19 survivors demonstrated some degree of fatigue or restriction in strenuous activity during 2–9-month follow-up.^51, 53 The Yorkshire Rehabilitation Scale (C19-YRS), Chalder Fatigue Scale (CFQ-11), Post-COVID-19 Functional Status Scale (PCFS), and Functional Impairment Checklist (FIC) items were used once in four different studies, reporting mild-to-moderate impairments across various domains of functioning for varying proportions of long-COVID individuals.^{29, 34, 45, 47}

Inflammatory markers

Seven studies evaluated inflammatory biomarkers, three of which evaluated their association with long-COVID in general,^27, 42, 46 and four examined their association with at least one pain domain.^{38, 39, 50, 56} Two studies found no significant differences in inflammatory biomarker profiles for long-COVID.^27, 46 While elevated levels of D-dimer and IL-6 and decreased lymphocytes during the acute phase of COVID-19, were associated with the persistence of muscle and joint pain,^39, 56 increased fibrinogen values were associated with the persistence of chest pain symptoms.³⁸ Elevated levels of CRP or ESR and D-dimer, during postacute phase COVID-19 follow-up, were significantly correlated with the occurrence of long-COVID-19 arthritis and pulmonary sequelae, respectively.^42, 56 While elevated levels of all five low-grade inflammation (LGI) composite indices, assessed at least 3 months after acute COVID-19 resolution, showed a positive correlation with long-COVID in general, elevated fibrinogen levels were reported to be significantly associated with persistent myalgia.⁵⁰

Factors associated with long-COVID

Twenty of the included 26 component studies reported at least one factor to be associated with increased incidence of any long-COVID symptom, 12 of which provided risk factor association with at least one persistent pain symptom (Table 2).

Being a female was reported to be associated with an increased risk of experiencing any symptom persistence past the acute phase of COVID-19 in five studies^{18, 28, 46–48} and for pain symptoms specifically, especially headaches, in two studies.^28, 55 Preexisting comorbidities associated with overall long-COVID symptom development were reported in six studies and included autoimmune diseases, depression or anxiety,⁴⁶ hypothyroidism,⁵² chronic rhinosinusitis,²⁷ hypertension and/or chronic kidney disease,⁴⁵ and presence of at least one medical comorbidity of those evaluated.^34, 47 In addition, comorbidity association with persisting pain symptoms specifically was reported in four studies and included increased BMI,^39, 49 any medical premorbid condition, and current smoking status.^39, 56, 57 Presenting at the acute phase of COVID-19 with various symptoms (eg, myalgia, headache, and arthralgia),^33, 43, 56 more severe clinical¹¹ and/or radiological^38, 39 manifestations or subsequently requiring a longer duration of hospitalization³⁹ were associated with pain symptoms persistence. However, one study reported that those who experienced the acute phase of COVID-19 more subtly, that is, did not require hospitalization or demonstrated any elevations in inflammatory biomarkers, had increased risk specifically for COVID-19 pain symptoms persistence.²⁸ Concerning any other long-COVID variant, increased risk was associated with requiring hospital admission in four studies,^{27, 34, 52, 55} more severe forms (eg, clinically, diagnostically, longer inpatient duration) in seven studies,^{11, 18, 27, 28, 34, 47, 52} and with headache in one study.⁴³

While one study reported that steroidal treatment during the acute phase of COVID-19 was associated with a significantly lower risk of suffering from headache, dysphagia, and chest pain, as well as leading to other long-COVID symptoms attenuation and QoL improvement,⁴¹ another study reported that the same intervention increased the risk of developing long-COVID symptoms such as fatigue or muscle weakness.⁴⁸ Moreover, being a health care worker,^27, 45, 47 police worker, housewife, and private sector job holder, as well as having rural geographical residence and Rh⁺ blood type,⁴⁷ were reported to increase the risk of developing any long-COVID symptom in general.

DISCUSSION

To the best of our knowledge, this is the first systematic review (SR) and MA to evaluate incidences for the entire repertoire of long-COVID’s potential persistent pain symptomatology. Various pain symptoms grouped under the chest, GI, musculoskeletal joint, musculoskeletal muscle, general body, and nervous system domains were established to be experienced up to one year by approximately 8%, 6%, 18%, 18%, 17%, and 12% of all individuals recovering from confirmed COVID-19, respectively. We noticed several factors to be associated with increased rates of general and/or pain-specific symptom persistency. Moreover, we also noticed that these long-term symptoms were associated with considerable levels of functional and QoL impairments.

Musculoskeletal pain symptoms dominance

Of all domains, musculoskeletal pain symptoms were the most commonly reported, which is in line with the original need to designate two specific and two related musculoskeletal domains (eg, joint, muscles, chest, and GBP, respectively). Moreover, musculoskeletal pain symptoms were previously reported to be among the most common and persistent of all long-COVID’s symptoms.^26, 75–77 It should be noted that these pooled proportions represent an amalgamation of various pain symptoms under domains representing their desired system-based allocation while also accounting for some possible regionally imposed anatomical limitations (Table 1). The latter accounts for cases in which “compromised” allocation was made based on provided inconclusive definitions (eg, not clearing the meaning of “chest” pain). Of all pain symptoms, musculoskeletal were previously reported as the most common to exist before as well as during the acute phase of the disease.³³ Thus, preexisting musculoskeletal pain exacerbation could account, to a certain degree, for its proportional dominance in long-COVID.^78–80

The proportional dominance of musculoskeletal pain symptoms persistence is especially intriguing when considering that these were specifically reported to demonstrate some opposite associations and clinical trends compared to those demonstrated by other commonly reported general long-COVID symptoms. Most notable were their increased associations with female gender, young age, and less severe acute course of COVID-19, as well as their directly related progressive temporal-proportional trend, rather than the usual subsiding nature of most other symptoms.²⁸ The former is in line with the reported increased risk of overall pain symptoms persistence being associated with patient factors, which, for the most part, reduce the odds of seeking medical attention (eg, mild symptoms, young, and healthy).³¹ Our GBP domain was intended to accommodate general and widespread pain symptoms. However, it turned out that differentiating these from the relatively “localized” musculoskeletal pain symptoms was more amenable than those of a more systemic and less defined nature (Table 1). It is much possible that the same holds as a basis for deciding to include these definitions from the begin with by some of the included studies in their questionnaires/assessment tools of long-COVID symptoms.^{28, 33, 45, 46, 51, 57} The relatively high and practically identical pooled proportions of the GBP domain (0.17) with those of the musculoskeletal (0.18 for joint and muscle groups), as well as its significant intergroup proportional difference in favor of the nonhospitalized group compared to the hospitalized (0.305 and 0.088, respectively; p_subgroup = 0.009; Table 3) reflect its kinship with these domains by highlighting their overlap in some of the abovementioned unique features, respectively.

Overlapping clinical conditions with long-COVID pain symptoms

Pain symptoms should be interpreted in the context of other general long-COVID symptoms. Of the more frequently, and to some extent concomitantly reported, are those of malaise, fatigue, dyspnea, and various cognitive and neuropsychiatric disorders.^30, 81, 82 The multifactorial and complex etiological nature, as well as the lack of specificity of these symptoms, underlie their challenging clinical assessment.⁸¹ Interestingly, fatigue was also one, if not the most common, and persistent symptom to be reported during the acute phase of COVID-19 and/or as part of its long-term sequela, respectively.^{11, 18, 28, 75, 77, 83} When carefully considering these aspects together with the abovementioned issues concerning the GBP and musculoskeletal pain domains, and the non-negligible proportions and clinical diversity demonstrated by the rest of the pain domains should reveal the existence of some overlapping features with previously established clinical entities.

Resemblance with postviral conditions

The general and nonspecific, as well as some of the organ-system-specific long-COVID symptoms, may be somewhat reminiscent of other postviral syndromes.¹¹ For example, resemblance could be found with the postinfectious fatigue syndrome, which was reported to follow Influenza as well as other coronaviruses such as the severe acute respiratory syndrome coronavirus (SARS) and Middle East respiratory syndrome coronavirus (MERS).^84–86 In addition, protracted neuro-cognitive and pulmonary impairments and symptoms,^84, 86–92 sleep disturbances,⁹³ myalgias,⁹³ and overall resulting functional disabilities,^{87, 88, 90–92} were previously reported in subsets of individuals who contracted SARS and/or MERS. Moreover, in line with previous reports of shared similarities,⁹⁴ a relatively recently published systematic review noticed that most long-COVID individuals report experiencing the major criteria symptoms of Myalgic Encephalomyelitis/Chronic Fatigue syndrome (ME/CFS), which also tends to follow various infectious agents.⁹³

Resemblance with fibromyalgia

Persistent pain symptoms of long-COVID, on the other hand, demonstrate resemblance with various features of other chronic pain conditions and syndromes,^{81, 82, 95–97} with some proposals arguing even for their shared clustering under the same group of syndromes.⁹⁶ Of these, intriguing similarities could be noticed with fibromyalgia (FM), and especially with its defining symptoms concerning the existence of persistent and multiple widespread painful foci throughout the body.⁹⁸ These, together with other characteristic symptoms such as fatigue, intestinal disorders, insomnia, and mood alterations, express a small part of the diverse range of possible FM manifestations able to overlap with those of long-COVID.^{82, 99, 100} Hypersensitization to painful stimuli in FM individuals is assumed to be triggered by various physical or emotional factors, though a lack of such triggers is not uncommon.¹⁰¹ These factors have been substantially associated with the multisystemic infection and inflammatory state of acute COVID-19, and can, therefore, precipitate a new-onset of “FM-like” symptoms (eg, pain and fatigue) or exacerbate already existing ones in FM patients.^31, 102 While a relatively exhaustive literature coverage of the latter was provided, the former has only recently begun to gain more attention with case reports and studies providing seminal evidence for a constellation of distinct features in line with the new-onset of FM, including young age at presentation, female sex, and new onset symptoms of widespread pain, fatigue, as well as a myriad of neuropsychiatric and sleep-related abnormalities.^{31, 103, 104} In line with this, female sex^28, 55 and young age²⁸ were reported to be associated with an increased risk of long-COVID pain symptoms by some of the included studies, though also increased age.^56, 57 The reported clear difference made by patients between their preexisting musculoskeletal pain symptoms with those of myalgia experienced during the acute phase of COVID-19 (defined by authors as “viral-induced generalized muscle pain”),³³ provides further support for the appropriateness of clinical distinction made for these FM-like widespread pain symptoms. Moreover, the high proportions demonstrated for the GBP domain specifically and musculoskeletal in general are in line with reports from a restricted number of studies indicating that during long-term follow-up periods, up to 20%–40% of all long-COVID cases qualified the criteria for FM diagnosis.^31, 105 Further resemblance is demonstrated by considerably low levels of specificity characterizing the individual clinical manifestations of these overall multifaceted conditions,^{93, 101, 106–108} both of which are yet to be fully understood, especially with regards to their underlying etiological agents, associated risk factors, and pathophysiological mechanisms.^11, 101 Thus, this compilation of findings may suggest that a potentially underestimated, yet possibly significant association exists between FM and long-COVID pain symptoms persistence.

Emerged just in recent years, COVID-19’s long-term consequences are still unraveling.^{1, 81, 83, 109} Therefore, a detailed discussion of shared mechanisms to underly the development of its remarkably protean chronic pain symptomatology, is not only beyond the scope of this work, but may even appear at first ostensibly farfetched. On the other hand, FM diagnosis, a relatively new entity on its own, was formally recognized during the last decades of the 20th century,¹⁰¹ which provides it with a much more solid scientific base and background compared to the former. To variable extents, the same can be argued in favor of the abovementioned postinfectious syndromes. Considering their just mentioned similarities and possible associations, it is not unreasonable to use some of the already established underlying clinical and pathophysiological principles of these long-COVID reminiscent syndromes as a deductive road map for gaining and complementing possible new insights regarding the mechanisms and associated risk factors of pain symptoms chronification in long-COVID, while keeping in mind that despite these similarities, there are unique aspects to this syndrome as well.⁸²

The multiple mechanisms suggested to underly long-COVID development reflect general, and nonspecific virulence factors of the SARS-CoV-2 infection frequently used to describe both acute and long-term consequences of COVID-19.^{30, 39, 83, 110} Nonetheless, new onset or exacerbation of chronic pain symptoms is most likely to occur via some of these mechanisms, which currently are neither pain-specific nor are they completely understood.^11, 81, 82 Grossly dichotomizing chronic pain symptoms to those originating from acute phase COVID-19 organ-specific damage or nonresolving long-lasting inflammatory mechanisms^11, 111 provides a comfortable conceptual framework for further discussion and is frequently being employed for discussing long-COVID development.⁸¹

Direct viral damage is usually more pronounced in the more dominantly expressing angiotensin-converting enzyme 2 (ACE2) tissues, the functional receptor of the SARS-CoV-2, which increases their susceptibility to its virulency-associated damage.¹¹² Of these, lung alveolar and small-intestine epithelial cells show the highest ACE2 expression levels and are also part of the body’s shared interface with the external environment, which accounts for their dominant role in viral transmission as well.¹¹³ Vascular endothelial and arterial smooth muscle cells throughout the entire body also demonstrate high ACE2 expression levels.¹¹⁴ This might represent their infrastructural role in facilitating the generation of widespread and nonspecific foci of tissue damage, through direct viral or indirect systemic inflammatory vascular effects,^{11, 115, 116} which accounts for some of the diversity seen in COVID-19 organ-system involvement.^117–119 Pathological innate immune system activation at viral uptake and replication areas, which leads to “site-specific” inflammation and vascular damage-induced hypoxia with subsequent tissue necrosis and organ-system damage,^{18, 115, 120} probably account for long-COVID symptom development.⁸¹ However, “systemic-sequalae” such as immune dysregulation with loss of self-tolerance and hyperinflammation, viral latency and reservoir persistency,¹¹⁸ endotheliopathy and thrombo-embolic/coagulative processes hyperactivation,^{46, 117, 121–123} and multiorgan and autonomic nervous system dysfunction⁹⁶ could further fuel previously mentioned pathological processes, leading to their overall chronification and amplification.¹¹⁹

Dysfunctional neuro-circuitry processing of pain signals, due to various neuro-transmissional related alterations, are assumed to represent the central pathophysiological drivers of FM’s development.¹²⁴ It is possible for similar alterations to follow the nonexclusive, though frequently reported, “site-specific” involvement of various nervous system elements in COVID-19 (eg, direct viral encephalitis, neuroinflammation, and cerebrovascular involvement).^{11, 26, 84, 86, 125} This could be further supported by FM’s reported association with viral diseases¹²⁶ and considerable similarities with long-COVID,⁹⁶ as well as COVID-19’s reported consequences of nervous system structural and functional changes.^127, 128 Although FM is considered a central sensitivity syndrome, pain sensitization is not a CNS-exclusively driven phenomena, rather it may well involve also peripheral and psychosocial causes for pain sensitization and chronification,^129, 130 reflecting the multidimensional aspects of pain perception.¹³¹ Peripheral pain generators in FM have been previously associated with peripheral structural and functional abnormalities, resulting in increased nociceptive tonic input, which feeds central pathways and may eventually lead to overt central pain sensitization.^132–134 COVID-19 site-specific damages of any possible organ system may well initiate similar peripheral pain-generating foci, though in a less musculoskeletal exclusive fashion, while its “systemic-sequalae” may maintain and perpetuate these long enough to qualify as chronic long-COVID pain symptoms. The former may also account for the considerable versatility demonstrated by the various system-based pain symptoms in our study, while the latter for their overall chronification. Selective attention allocation, being dedicated to external and internal pain-related signals, is believed to represent the cognitive-emotional drivers of pain sensitization in FM individuals.¹²⁹ Neuropsychiatric vulnerability^{11, 46, 135} and pre-existing chronic pain symptoms exacerbations^{29, 31, 36, 49} reported to be associated with increased risk of COVID-19 pain symptoms persistence, could to some extent account for these cognitive-emotional aspects. However, among the commonly reported neurological symptoms of long-COVID are those of general and nonspecific nature, such as decreased concentration, brain fog, memory impairments, and various other cognitive impairments.^13, 30 These most likely represent COVID-19 related diffuse neural network level impairments, rather than isolated lesions, caused by foci of “site-specific” involvement scattered across the CNS in a diffuse fashion.^136, 137 It is much possible that these nonspecific insults may also account for the development of dysfunctional overall attention allocation to painful stimuli in these patients. Lastly, interpersonal sensitization processes of neural networks involved in the contextual experience of pain within a certain social-group, represent the social dimension of pain, and may also be considered as a possible driver of pain sensitization in FM individuals.^138, 139 These may account, to some extent, the reports of increased pain burden being associated with social isolation restrictions and overall quarantine imposed during peaks of viral outbreaks, especially in the prevaccine era of COVID-19.^18, 140

Long-COVID pain symptoms inflammatory correlates

The contribution of various cellular and molecular components involved in pro-inflammatory pathways for chronic pain development, via both central and peripheral mechanisms, is well-established.^141–143 Of the four studies to explore inflammatory parameters specifically for pain symptoms persistency development, significant associations were reported by all four subset of individuals continuing to present with increased levels of inflammatory biomarkers past the acute phase of COVID-19.^{38, 39, 50, 56} Similar reports of increased inflammatory biomarker profile association with various postinfectious syndromes were previously made.^28, 144 These reports are in line with the non-specific nature of the more commonly tested “general” inflammatory parameters, which accounts for the relative ease in designating them as risk markers as opposed to risk factors, which necessitates demonstrating causality rather than just statistical association.^145–147 Nonetheless, it is much possible, that persistent inflammation accounts to some extent of pain symptoms persistency in COVID-19, and was previously proposed by various studies.⁸¹ It seems that dysregulated immune response, attributed to the interplay between the patient’s overall risk factor profile with COVID-19, underlies the subsequent possible maladaptive activation patterns of various inflammatory mediators and pathways.^81, 82

Interestingly, low-grade inflammation (LGI) composite indices were evaluated in one component study, which reported a positive association between chronically increased values and pain symptoms persistency.⁵⁰ This association was previously reported for various organ system-related persistent symptoms in patients recovering from COVID-19.⁸¹ LGI is a term used to describe subclinical states of mildly elevated inflammatory biomarker profile, that have not been clearly defined yet.¹⁴⁸ However, it was associated with various chronic conditions such as malignancies, neurodegenerative and metabolic disorders, and cardiovascular diseases (CVDs) and even was recognized as an independent risk factor for some.^149, 150 Atherosclerosis is the underlying pathological process of the major vasculopathic conditions, involving large and medium-sized arteries collectively referred to as CVDs (eg, PAD, coronary heart disease, and cerebrovascular disease). The consistent presence of various components and pathways of the innate and adaptive immune response, falling under the broad term of “inflammation”, throughout all phases of the atherosclerotic process has been well-established.^151–153 These, in essence, translate the many well-established risk factors of CVDs (eg, HTN, dyslipidemia, smoking) into atherosclerotic tissue damage, which also led to their current recognition as chronic immune-mediated inflammatory diseases.^{151, 152, 154, 155} Similarly, it is much possible, that SARS-CoV-2 tropism for endothelial and immune cellular elements¹¹³ to underlie its role as a vasculopathic and immuno-dysregulatory risk factor. Moreover, the possible synergistic consequences of these factors, translated via inflammatory mechanism, may underlie acute and possibly long-term sequelae of COVID-19 in general, and of chronic pain genesis specifically.^81, 82 Parallel premorbid vulnerability of vascular infrastructure in CVDs, as well as their just mentioned association with dysregulatory immuno-inflammatory pathways,¹⁵⁶ could account for their reported increased risk of suffering more clinically severe and possibly prolonged COVID-19 symptoms of any kind.¹¹⁹ Similarly to its role in CVDs,^149, 150 LGI, could thus serve as possible risk factor of COVID-19 severity, by tonically providing pro-inflammatory “substrates,” which feed and augment, local and systemic pathological processes resulting from SARS-CoV-2 infection.

Although premorbid levels have not been studied, LGI states were reported to be associated with COVID-19 symptom persistency,⁸¹ including those of pain.^39, 50 Attributing causative properties and independent risk predictive capabilities, though yet to be established, could be of great value since LGI is known for both being easily detected by standard and simple assays as well as intervention responsive (eg, medical or lifestyle modifications).¹⁵⁰ While the former justifies its clinical utility,^157, 158 the latter emphasizes its importance as a possible target for disease prevention.¹⁵⁹

Long-COVID pain symptoms persistence associated factors

Sex

COVID-19 symptoms persistency in general^{18, 28, 46–48} and of pain specifically,^28, 55 were commonly reported to co-occur with female sex, which was also associated with increased duration of symptoms and functional limitations compared to males.⁴⁷ Previous studies have also mentioned these associations frequently,^{14, 160–166} though nonconsistently, as some (including in current study) reported similar or decreased risk compared to males.^{39, 163, 167, 168} Increased levels of anxiety and depression, altered behavioral reactions to pain, differences in central pain processing, and menstrual-related hormonal effects are some of the reported pathophysiological drivers to account for the increased prevalence of chronic pain disorders seen in females.^{106, 169, 170} It is much possible that these would also, to some extent, underlie the female sex as a risk factor for post-COVID-19 pain chronification.

Age

The association between increased age with persistency of general^28, 45, 48 and pain^56, 57 symptoms was reported and is in line with previous reports.¹⁶⁵ However, young age was also associated with symptom persistency, especially with those of chronic pain.^28, 47 This association was reported in previous studies that also noted its independence of acute phase severity,⁸² and may, to some extent, be accounted for their overall less symptomatic disease course and better baseline health status.^28, 82 These, as previously mentioned, contribute to the overall discouragement of pursuing medical attention by this age group.³¹ Increased age, on the other hand, was associated more frequently, though nonpain-specifically, with overall long-COVID symptom persistence, which might indicate different underlying pathomechanistic pathways from those in the younger population.^{28, 81, 171} Moreover, increased age is one of the recognized demographic risk factors to be associated with chronic pain development, as is the parallelly accompanying increase of overall morbidity burden see in this population.¹⁶⁹ It is reasonable to assume that these factors contribute to overall increased frailty and consequent vulnerability to COVID-19 effects, in much similar fashion described previously for the CVDs. This is somewhat in line with the various preexisting comorbidities frequently reported in subset of individuals continuing to experience persistent symptoms (including pain) past the acute phase of COVID-19.^{27, 34, 39, 45–47, 49, 52, 56, 57} Moreover, similar reports concerning general symptom persistence were also made in previous studies,^164–166 which overall enables us to assume increased risk association, however, not determining it for these factors.

COVID-19 acute phase severity indicators

Other frequently reported factors associated with symptoms persistency were various indicators of acute phase disease severity. Inconsistent reports concerning these were previously made, with studies arguing both for lack^{164, 166, 172–174} and presence^{94, 160, 175} of this association. Mentioned factors reported to be associated were the need for longer hospitalization, aggressive interventions at admission (eg, mechanical ventilation), and ICU admission^{94, 160, 175} as well as the increased burden of presenting symptoms.¹⁶⁸ Despite inconsistencies in clinical parameters used to represent initial disease severity in our included studies, overall similar associations were made with both general^{11, 18, 27, 28, 34, 47, 52, 55} and pain-specific^{11, 28, 33, 38, 39, 43, 56} symptoms persistency. These disease severity indicators are in essence representatives of the extent to which organ-system tissue damage is being translated into clinical expression. Therefore, considerable tissue damage could be expected in those demonstrating clinical need and findings justifying longer and more aggressive overall care, as well as to underlie the development of acute and long-term symptomatology of COVID-19. Supporting this association are various long-term physical and mental sequelae, which collectively constitute the postintensive care syndrome (PICS). These symptoms are known consequences of the characteristically extensive tissue damage seen in ICU-admitted patients surviving their critical illnesses.^176–178 Moreover, COVID-19 survivors, especially most severe cases, are known to be at increased risk of being admitted to ICU care, and consequently to suffer from PICS.¹⁷⁹ While this further supports the abovementioned association, it also obscures the independently additive effects which COVID-19 pathomechanistic factors might have, on top of those of PICS, on chronic symptom development and warrants further investigation.^179, 180

LIMITATIONS

Our SR and MA have several limitations, which should be kept in mind upon interpreting the received results. Identifying key methodological issues and pitfalls in included component studies and their interaction with the received results of the current study is important not only for their better understanding but rather also for highlighting their underlying role in core research issues concerning long-term symptom persistency in COVID-19.

Component study designs

A large proportion of nonprospective study designs precludes an accurate understanding of the temporal relation between the acute phase of COVID-19 with long-term symptoms (including pain) development and their clinical course. The observational nature of included studies enables, at best, to assume association but not causality. The latter is further undermined by the fact that most component studies reported about primary and/or secondary outcomes failed to corroborate their presence before acute COVID-19 episode and subsequently also their potential worsening. Long-COVID and its chronic pain symptoms were previously reported to be associated with other medical or mental sequelae of acute disease (eg, mood-affective disorders) or various repercussions arising from the socioeconomic turmoil created by the SARS-CoV-2 pandemic (eg, imposed social or economic activity restrictions).^{28, 164, 165} Yet, the vast majority of included component studies failed to include a control group of nonexposed individuals (Appendix S1), thus providing another barrier for inferring exposure-outcome relationship causality. Failing to reduce heterogeneity levels via subsequent stratification by study design can substantially be attributed to other variability generators not accounted by this subgroup alone, some of which are discussed in the next sections. Various aspects of long-COVID pain symptoms could be well addressed by various study designs. However, these should be conducted in line with appropriate quality guidelines (ie, NOS component requirements) to ensure the translation of these results to the intended target population.

Cohort selection

Several factors concerning the underrepresentation of specific populations were frequently noted and are likely to undermine the ability to generalize the results from this study for them. Of the most obvious were the large proportions of exclusively hospitalized cohorts together with a complementary paucity of asymptomatic community-based cohorts. Moreover, the former might account for considerable selection bias, as these individuals tend to be older, suffer from more comorbidities, and face higher rates of hospital-related conditions such as PICS, all of which could independently account for increased risk of chronic pain development.^169, 179 Moreover, the few studies that did include both populations unanimously addressed them as singular cohorts without providing adequate stratification to enable their separate assessment. Even when examining the few nonhospitalized cohorts, it is still evident that most of these are based on individuals that actively sought medical attention (eg, reached outpatient or community-based clinics). The latter probably reflects certain above-threshold levels of overall symptom burden being experienced, which may further imply the underrepresentation of asymptomatic individuals. Considering that only RT-PCR and/or SARS-CoV-2 antibody confirmed cases were included by component studies suggests underrepresentation of those suffering from long-COVID-19 symptoms but were not confirmed due to various reasons, such as but not limited to assay sensitivity, inadequate tissue sampling, and asynchrony with viral shedding or antibody production.⁴⁶ Thus, it is vital that future studies clearly define their target population and provide them with adequate control groups. These will facilitate better overall generalizability of the results and the ability to infer conclusions from the received results.

Follow-up period

Considerable variation was noticed in follow-up periods and definitions used to consider patients as suffering from long-term symptoms across all studies, which further obscures temporal and casual associations. Generalizing results from this study is also problematic for SARS-CoV-2, specifically vaccinated or treated individuals, as none of the component studies included them in their cohorts. These constitute an important knowledge gap that should be addressed in future studies when sufficiently enough long-term follow-ups, including these populations, will accumulate.

Outcome ascertainment

Another important limitation was the considerable use of nonstandardized assessment tools for evaluating overall symptom persistency, including pain (eg, symptom questionnaires), with none of the studies providing reports of any validity and reliability testing for these instruments. For the most part, items of pain symptoms were included as part of general nonspecific symptom lists, which may contribute to recall bias. Vocabular diversity used in describing apparently similar clinical manifestations among component studies further supports recall bias and represents another lack of standardization aspect. Active and passive inquiry methods for symptoms ascertainment applied by various studies could explain some of the considerable range demonstrated for apparently similar symptoms, as it was previously reported that passive methods tend to produce lower incidence rates than active ones.¹⁸ These issues are reinforced by findings of a recently conducted study, which explored the various terms used to described different long-COVID symptoms, including those of pain, in 59 overall studies. Authors have reported considerable variation in apparently synonymous terms, which were used to describe the same apparent symptoms. Myalgia, for example, was described also using the terms myalgia/persistent muscle pain, myalgias-arthralgias, muscle or body aches, muscle/joint pain, muscle/joint pain, flulike symptoms, muscle pain, and muscle aches. This variability was considerably attributed to the relative novelty of long-COVID, lack of census definitions, and considerable differences in terminology used to describe patient-reported symptoms.¹⁸¹

Thus, asides from considering the adoption of a rather underestimating approach for symptom inclusion in our study (methods section), any inference regarding their proportional values is somewhat difficult due to previously reported wide ranges.^31, 33, 81 However, when considering the overall variability just mentioned in the terms and follow-up periods used, it might be that the calculated pooled proportions are incompletely accurate. Moreover, it is also possible that the selected keywords used in our search strategy did not provide for extensive enough framework to encompass all possible terms used for describing chronic pain symptoms throughout previously published literature, leading to their potential misrepresentation.

High levels of heterogeneity

Another limitation to consider is the substantial levels of calculated heterogeneity constantly demonstrated across all our meta-analyses, which could be attributed to the just mentioned significant methodological variability among our component studies. Methodological variation was previously recognized for its possible contribution to the inconsistent rates of long-COVID symptoms being reported across the literature.^{81, 182–184} Moreover, pooling together proportions produced by studies that use inconsistent definitions for outcome variables, different cohort characteristics, and various methods for data ascertainment is a known pitfall responsible for the characteristically high heterogeneity of proportional meta-analyses in general. These are known for potentially introducing many biases and confounding factors (often being difficult to quantify), which are also left “unaccounted” considering the nonrandomized design of observational studies.¹⁸⁵ These could also explain why significant heterogeneity reduction was not achieved even after stratifying by certain subgroups.

The considerable underlying variability of the abovementioned limitations seems to be the most constant motif across all sections of this study. Ad hoc critically analyzing and discussing the results of this study provided possible explanations for the overall high heterogeneity levels of our attempts to better characterize the long-term persistence of pain symptoms in those recovering from COVID-19. However, this heterogeneity may also be attributed to or even overtly accounted for the fact that these long-term symptoms, which for the most part are still being ascribed under a single entity, are in fact a multitude of diverse clinical conditions stemming from heavily intertwined unique and/or shared pathomechanistic pathways or associated risk factors. Initial steps in consolidating this notion were previously provided by several studies reporting more granular levels of experienced symptomatology by describing several distinct phenotypes¹⁸⁶ or patterns⁷⁷ of long-term clinical sequelae. To facilitate a more accurate description of these possible various long-term clinical conditions, future studies should address long-COVID symptoms specifically and comprehensively rather than aggregately and superficially while implementing widely accepted consensus definitions. Altogether, these limitation, and lack of variability reduction upon the various subgroup analyses, should highlight the need for a more holistic approach for the multifaceted restructuring of the current study designs components upon conceiving new ones, which will be aimed from their outset to address pain symptoms specifically.

CONCLUSIONS

Herein, we determined that non-negligible proportions of acute COVID-19 survivors continue to experience persistent chest, gastrointestinal, musculoskeletal joint and/or muscle, general body, and nervous system-related pain symptoms. Persistently considerable heterogeneity levels were demonstrated across all these domains and in their subsequent subgroup analysis attempts, which limits our further understanding of these various long-term clinical conditions. Realizing the latter requires future research efforts to reduce variability generators by addressing long-COVID pain symptoms specifically and comprehensively by examining more granular levels of clinical expression, using standardized and validated assessment methods, implementing narrow consensus definitions, and developing adequate study designs, which will enable the wide generalization of their results.

Introduction

Objectives

Methods

Results

Conclusion

INTRODUCTION

METHODS

Literature search and reporting

Guidelines and selection criteria

Data extraction and outcome measures

Quality of evidence and risk of bias assessment

Data synthesis and analysis

RESULTS

Study selection and search results

Study characteristics

Quality of evidence and risk of bias assessment

Results synthesis

Chest pain symptoms

Gastrointestinal system pain symptoms

Musculoskeletal system joint pain symptoms

Musculoskeletal system muscle pain symptoms

General body pain symptoms

Nervous system pain symptoms

Heterogeneity

Publication bias

Functional outcomes and quality of life

Inflammatory markers

Factors associated with long-COVID

DISCUSSION

Musculoskeletal pain symptoms dominance

Overlapping clinical conditions with long-COVID pain symptoms

Resemblance with postviral conditions

Resemblance with fibromyalgia

Long-COVID pain symptoms inflammatory correlates

Long-COVID pain symptoms persistence associated factors

Sex

Age

COVID-19 acute phase severity indicators

LIMITATIONS

Component study designs

Cohort selection

Follow-up period

Outcome ascertainment

High levels of heterogeneity

CONCLUSIONS

Leave a Reply Cancel reply